Cross-Comparison Between Jilin-1GF03B and Sentinel-2 Multi-Spectral Measurements and Phenological Monitors

• Published in: IEEE access Vol. 12; pp. 43540 - 43551
• Main Authors: Sun, Jiayue, Zhu, Ruifei, Gong, Jialong, Qu, Chunmei, Guo, Fengxiang
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Cloud cover; Correlation coefficients; Earth Observing System; Earth surface; Indexes; Jilin-1; Land use; Performance evaluation; phenological variations; Phenology; radiation consistency; Remote observing; Remote sensing; Remote sensors; Satellite broadcasting; Satellite constellations; Satellite images; Satellites; Sensitivity; Sensitivity analysis; Sensors; Time series analysis; time-series; Vegetation index; vegetation indices; Vegetation mapping
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3379322

The necessity of combining multi-source remote sensing sensors for multi-temporal observation of the Earth, arises from the influence of cloud cover and the limited operational period of single satellite. However, significant differences exist among satellites and among sensors within the same satellite system. In this case, conducting a cross-comparison analysis is crucial for reliable change detection. The Jilin-1 satellite constellation has a total of 108 satellites in orbit, forming the largest commercial remote sensing satellites constellation in China. Firstly, the top of atmosphere (TOA) data from the Jilin-1GF03B series satellites (Jilin-1GF03B04, Jilin-1GF03B05, and Jilin-1GF03B06), following radiometric calibration processing, were selected to evaluate their performance in a single phase by comparison to Sentinel-2 satellites. Subsequently, we compared four spectrum bands and four commonly used biophysical indices, and the results indicated a higher radiation consistency between Jilin-1GF03B and Sentinel-2 satellites, suggested by correlation coefficient (over 0.953), and R2 (over 0.907). Moreover, we assessed the band sensitivity of the two satellites across different land use classes. It was found that Sentinel-2 exhibited greater sensitivity in bare land, vegetation, and forest, whereas the near-infrared band of Jilin-1GF03B proved more sensitive in water areas. Finally, to mitigate the risk of randomness in a single temporal phase, we constructed two long-term vegetation index datasets for phenological analysis. The close performance alignment observed for both single-phase and phenological variations indicated that Jilin-1GF03B satellites and Sentinel-2 can be effectively utilized together for long-term monitoring of Earth surface attributes.